Sa1916 RNA-Sequencing Approach for the Identification of Novel Long Non-Coding RNA Biomarkers in Colorectal Cancer

Sa1916 RNA-Sequencing Approach for the Identification of Novel Long Non-Coding RNA Biomarkers in Colorectal Cancer

AGA Abstracts plotted along each chromosome. 7. Mutations in the TGF-β (transforming growth factorβ) and CEA pathway members were observed in 5 out o...

66KB Sizes 0 Downloads 24 Views

AGA Abstracts

plotted along each chromosome. 7. Mutations in the TGF-β (transforming growth factorβ) and CEA pathway members were observed in 5 out of 11 adenomas, overlapping with wnt/p53 mutations in 4 of adenomas: All 4 were TVAs. 8. Further analyses of expression levels of CEA and TGF-β pathway members in 30 non dysplastic adenomas and normal colon tissues revealed a marked increase (over 8 fold) in CEA expression in 25% of adenoma samples which was linked to concomitant loss of TGF-β signaling. 9. Functional studies revealed CEA association with the TGF-β Type I receptor and disruption of TGF-β tumor suppressor signaling with activation of STAT3. Conclusions: Small adenomas both TVAs and SSAs can resemble CRCs in genomic profiling and may reflect a high risk population. Disruption of the CEA/TGF-β pathway in early adenomas may reflect a new and early role for these pathways in CRC. The molecular signatures we characterize here support new approaches to biomarker driven targeting of CEA/TGF-β in high risk adenomas, potentially improving survival of CRCs.

Sa1917 Urine-Based Metabolomic Diagnostic Tests Predict Response to Chemotherapy and Adverse Events in Patients With Colorectal Cance Mark Dykstra, Noah J. Switzer, Roman Eisner, Victor Tso, Rae R. Foshaug, Richard N. Fedorak, Haili Wang Background: Colorectal cancer is the third leading cause of cancer-related death in the western world. The ability to predict an individual patient's response to particular chemotherapy regimens would be of great value for clinicians and patients when planning cancer treatment. Urine metabolomics may be that diagnostic tool. It provides a non-invasive snapshot of a patient's state of health by examining the presence and concentration of individual metabolites creating a unique, patient-specific metabolomic fingerprint. Aim: To develop a metabolomics-based biomarker panel to identify adverse events and response to adjuvant chemotherapy following resection for stage III to IV colorectal cancer. Methods: We performed a retrospective chart review of patients with diagnosed with stage III or IV colorectal cancer presenting to four tertiary care hospitals in Edmonton between 2008 and 2012. Exclusion criteria included chemotherapy for palliation of colorectal cancer and patients living outside of Alberta. Data was collected for chemotherapy regimen, adverse events associated with chemotherapy, and disease progression and recurrence. Chemotherapy adverse events were subdivided into delays in treatment and chemotherapy dose reductions. Patients provided urine samples for metabolomic analysis prior to neoadjuvant treatment or surgical resection. One-dimensional nuclear magnetic resonance (NMR) spectra of urine samples were acquired using an Oxford 600 MHz NMR spectrometer with a Varian VNMRS two-channel console. The 1H NMR spectrum of each urine sample was analyzed using Chenomx NMRSuite v7.0 (Chenomx Inc, Edmonton). Using machine learning, a predictive biomarker panel was generated and evaluated using 10-fold cross-validation. Results: Patients fulfilling the inclusion criteria were grouped by cancer stage determined at the time of resection: stage 1, n=3; stage II, n=4; stage III, n=49; and stage IV, n=4. Urine spectra were obtained for each of the 60 patients. The area under the ROC curve for our metabolomicbased biomarker panel relative to adverse events was 0.750 for treatment delay and 0.542 for dose reduction. Pertaining to disease progression, the area under the ROC curve for cancer stage was 0.725, 5 year survival was 0.675, and recurrence 0.650. Conclusions: We have developed a metabolomics-based biomarker fingerprint to predict which colorectal cancer patients who will or will not have an adverse event to chemotherapy curative resection and a fingerprint to predict cancer stage, 5 year survival, and disease recurrence. We are currently refining these panels to address other aspects of patient response to adjuvant chemotherapy

Sa1915 Chromosome Translocation T(2,10,16) Marks ‘Point of No-Return' in Barrett's Epithelium Carcinogenesis, In-Vitro Model Manisha Bajpai, Hana Aviv, Kiron M. Das Barrett's epithelium (BE) is a sequel of inflammation resulting from chronic gastroesophageal reflux and is a major risk factor for esophageal adenocarcinoma (EAC). Molecular events in BE pathogenesis are still unclear and therefore, stratification of patients at higher-risk for EA remains a clinical challenge. To gain insight into the molecular processes, we developed a novel in-vitro BE carcinogenesis (BEC) model (Int J Cancer 2011). Benign, hTERT immortalized BE cells (BAR-T) were exposed to acid and bile at pH4 (B4) 5 min each day for 60 weeks, as they accumulated progressive molecular and phenotype changes (in weeks, W) e.g., amplification of colonic / intestinal phenotype mAbDas-1 and CK8 (Lab Invest. 2008), chromosomal aberrations (BEC20W) (Mol. Cytogenet. 2012), change in cell shape with clumping (BEC40W) soft agar colony formation (BEC60W) and finally, tumors in nude mice indicative of malignant transformation (Int J Cancer, 2011). A parallel set of cells not exposed to B4 did not show these changes. Aim: Identification of a molecular event(s) that marks the ‘point of no return' or higher risk of un-impeded progression to neoplasia in the BEC model. Methods: BEC20W and 40W cells were revived from storage and divided into two groups, only one group was exposed to B4, 5 mins everyday and compared to the untreated group. After 20wks, cells were assessed for following endpoints- cell shape and clumping, colony formation in soft agar. Chromosomal changes were further investigated by cytogenetic analysis after chromosome painting. Results: Only BEC20W cells exposed to B4 for additional 14 more weeks changed to oval shape with characteristic clumping (i.e. BEC34W). The BEC40W cells with or without further exposure to B4, retained the oval shape and acquired the ability to form colonies in soft agar at the end of the 20 weeks observation. Chromosome painting of replicates of BEC34W cells (BEC20W+14wks B4) confirmed recurrent appearance (in six out of eight replicates) of a chromosomal translocation t(2,10,16)(p23,q22,q22). Conclusion: The BEC model demonstrates that absence of B4 trigger could prevent progression of neoplastic changes in BEC20W cells but not in BEC40W cells that appear to be committed to progressive transformation i.e., "point of no return". This study identified a novel translocation t(2,10,16) [involving chromosome segments harboring known carcinogenic genes] that reliably appears in the BEC34W cells that are just six weeks away from the ‘point of no return' (BEC40W). Translocations are considered biomarker events in leukemia, prostate and lung cancers, but have not been investigated in BE / EAC. The t(2,10,16) after clinical validation may be useful as a fluorescent in-situ hybridization (FISH) based predictive marker to stratify BE patients at higher risk of developing EAC.

Sa1918 TGR5, a Bile Acid Receptor, Is Highly Expressed in Esophageal Adenocarcinoma and Precancerous Lesions Significantly Associated With Worse Overall Survival and Gender Difference Chunhong Pang, Amy A. LaLonde, Tony E. Godfrey, Sayak Ghatak, David A. Dean, Jun Sun, Jiping Da, Tongtong Wu, Zhongren Zhou Background: Bile acid reflux to esophagus plays an important role in carcinogenesis of esophageal adenocarcinoma (EAC). Recently, TGR5, a G-protein coupled bile acid receptor, is reported to associate with the development of several types of cancer, such as esophagus, stomach and liver. However, little is known about the distribution of TGR5 expression in human esophageal adenocarcinoma and precancerous lesions. Methods: TGR5 expression was analyzed by immunohistochemistry from tissue microarrays (TMA), containing 106 EAC, 35 Barrett's esophagus (BE), 18 low-grade dysplasia (LGD), 11 high-grade dysplasia (HGD), 62 columnar cell metaplasia (CM) and 83 squamous mucosa (SE). Patients' survival data, histological diagnoses and tumor staging were collected. The intensity (0-3) and percentage of the TGR5 expression in TMA slides were scored. In an independent series of 116 esophageal adenocarcinoma samples TGR5 amplification was analyzed using Affymetrix SNP 6.0 microarrays. Results: TGR5 protein was highly expressed in 71% EAC (75/106), 100% HGD (11/11), 72% LGD (13/18), 66% (23/35) BE, 84% (52/62) CM and 36% (30/ 83) SE group. TGR5 high expression in the columnar lesions is significantly greater than that in SE group. Patients with TGR5 high expression was significantly associated with worse overall survival comparing to those with non-high-expression (p=0.0432). TGR5 high expression was significant higher in males than in females in all cases with the odd of 1.9 times. In addition, we found that TGR5 high expression was significantly related (P<0.0001) with VDR high expression, another bile salt receptor. TGR5 gene amplification was found in in 12.7% (14/116) esophageal adenocarcinoma, which is not associated with overall survival or disease-free survival. Conclusions: Our findings indicated that TGR5 might play an important role on the development and prognosis of esophageal adenocarcinoma through a bile acid ligand. Gender difference in TGR5 and VDR expression may help to explain why male has high incidence of esophageal adenocarcinoma.

Sa1916 RNA-Sequencing Approach for the Identification of Novel Long Non-Coding RNA Biomarkers in Colorectal Cancer Atsushi Yamada, Yoshinaga Okugawa, C. R. Boland, Ajay Goel Objectives: Long non-coding RNAs (lncRNAs) are transcribed RNA molecules, >200nt in length, which do not encode for proteins. Colorectal cancer (CRC) is the third leading cause of cancer-related deaths in the United States. Although a subset of lncRNAs are known to associate with human cancers, till date majority of lncRNAs have not been well characterized and the dysregulation of specific lncRNAs has not been fully explored in CRC. The aim of this study was to perform a systematic and comprehensive identification of novel lncRNAs dysregulated in CRC through next generation RNA sequencing technology (RNA-seq). Methods: Three pairs of matched CRC and adjacent normal mucosa (NM) tissues were utilized for RNA-seq. RNA-seq was performed by Illumina HiSeq 2000 platform. Expression levels of lncRNAs were analyzed by DESeq and edgeR, and a pairwise comparison between matched CRC and NM was performed using the GLM model. LncRNAs with a p-value ≤0.01 were selected as significantly dysregulated lncRNAs, and their expression levels were validated by quantitative RT-PCR assays in a cohort of 141 matched CRC and NM tissues. Results: A heatmap generated from RNA-seq data showed distinct expression patterns of lncRNAs between CRC and NM tissues. Eight lncRNAs were found to be significantly dysregulated in CRC compared to NM tissues, including 4 up-regulated and 4 down-regulated lncRNAs. Seven of them were novel, while one down-regulated lncRNA, MEG3, has been previously reported to be down-regulated in other types of human cancers. Dysregulated expression of lncRNAs found by RNA-seq was successfully validated in 141 pairs of matched CRC and NM tissues by real-time RT-PCR. Conclusions: Using RNA-seq technology, we identified novel lncRNA biomarkers which were dysregulated in CRC compared to NM tissues. These lncRNAs may play roles in colorectal carcinogenesis and possibly serve as therapeutic targets and/or diagnostic biomarkers for CRC. Our study highlights the use of RNA-seq as a powerful tool for the identification of novel lncRNAs which play important roles in CRC.

AGA Abstracts

Sa1919 Regulation of G1/S Checkpoint by Insulin Receptor Substrate-4 in Colon Cancer Cells and Its Possible Involvement in Colorectal Cancer Patricia Sanmartín-Salinas, Miguel A. Toro, Maria Chaparro, Borja Hernandez-Breijo, D. Cano-Martínez, Irene de los Dolores Román Curto, Javier P. Gisbert, Antonio Jiménez, Fernando Noguerales, Luis Guijarro BACKGROUND AND AIMS: Colorectal cancer (CRC) is the second leading cause of cancerrelated death and there are approximately one million new cases diagnosed worldwide each year. New biomarkers and therapeutic targets are needed for the diagnosis and treatment of the disease. MicroRNA 145 is a tumor suppressor of colon cells that targets insulin-like growth factor receptor and insulin receptor substrate-1. Our group showed that insulin receptor substrate-4 (IRS-4) also regulates this signaling cascade. In this work, we evaluate the possible implication of IRS-4 in colon cancer. METHODS: The studies were carried out both in vitro (RKO colon cancer cells) and in vivo (human colorectal cancer samples). To

S-354